- Town set attitude for affecting single creature, adapted from Windows code; technically redundant, but maybe handy
- If numeric response
- Print nums (for debugging)
- SDF arithmetic - add, subtract, multiply, divide, exponentiate
- Store random to SDF (adapted from Windows code)
- Display picture (inspired by Windows code, but the implementation is completely different and totally incompatible)
- Remove all field booleans except quickfire and belt, which have been moved to cCurTown
- Alter and extend place_spell_pattern, to allow arbitrary damage types and to make it more clear in the code what's happening when it's called
- Delete fields.cpp file; a few things moved to locutils.cpp, but most are now part of cCurTown
- set_terrain function automatically updates belt present boolean if setting to a conveyor.
- Includes new status effect images for the forcecage and for hypothetical inverses of dumbfound and magic resistance, as well as icons for the whole-party statuses.
Also some bugfixes and stuff:
- Fix specials sometimes being run twice in a row
- Holding Control while clicking Create also makes a debug party (as an alternative to holding Command)
- Fix "How Many" popup being non-dismissible
- Reduce loading time spent on checking for missing opcodes
Adapted from *i:
- Show a confirm dialog when interrupting a special node sequence
- New monster special ability: call global special node (as an action, not on death)
- New item special ability: call global special node
- Check there's a monster death special before calling it (wasn't necessary before, might be now with the special queue changes)
- Queue specials that are triggered while another special is in progress, instead of ignoring them; they will be run after the current special in progress finishes.
- *i's version of petrification touch is currently active only for monster-on-monster combat; need to merge with my version for monster-on-pc combat.
- Pass party location to special in use special item context
- Fix set town visibility node (was checking wrong field and thus could not hide towns)
Special nodes:
- Town Hostile: change to Set Town Attitude
- Select PC node: option to select random PC
- Affect special nodes can now affect monsters
- Fix affect death node reviving non-existent PCs
- Affect Spells: Can remove spells, and can affect level 1-3 spells
- If Objects: Merged from If Barrels and If Crates
- If Species: Replaces If Cave Lore
- If Trait: Replaces If Woodsman
- If Statistic: Replaces If Enough Mage Lore
- Change Lighting: Can affect town's global lighting setting, player's light level, or both at once.
- Pointers! Actually, I'd already implemented the callbacks for setting and getting them, but they're now actually used, and the implementation has been tweaked a little.
- Campaign flags! Again, I'd already implemented them sorta, but I tweaked things and they ended up sort of halfway between the two implementations. Plus there's now a special node to set them.
Additional bits:
- Special queue now uses an std::queue instead of a basic array.
- Enum for town lighting levels
- Disease touch ability is now honoured for monster-on-monster combat
- See monster special context now passes the monster's location as the trigger location; also, removed the double-trigger from one circumstance.
- Along with the set town attitude change, there's now the possibility for making the town hostile to trigger a special node, which can cause the party to be slain.
- Select PC special node: option to select specific PC
- Spell IDs for use in shops and Affect Spell nodes have changed so that 0 is now the first level 1 spell, and so forth.
- add_string_to_buf can now auto-split the string over multiple lines, and the special node that uses it takes advantage of this
- Special node parser warns if a node type is missing a corresponding opcode
- Reserved "pointers" to access the special node's trigger location (this was *i's idea, but he never implemented it)
- LED groups now trigger their own click handler in addition to the clicked LED's click handler (provided the latter returns true). If the handler returns false, the click has no effect.
- LED groups now cancel the selection change if their focus handler returns false; this mimics the behaviour when an individual LED's focus handler returns false.
- Move the dialog getResult() definitions inline - since there's only two of them now that I'm using boost::any, having them in a separate file is pointless.
- Changed how the pict choice dialog returns its result - now it returns only whether the user clicked cancel and provides getters to obtain the number and type.
- Pict and string choice dialogs now hide the arrow buttons when there is only one page of options.
- Fix pict choice dialog always returning the initially selected value (similar to how the string choice dialog did before I fixed it)
- When passed an invalid starting selection, the pict choice dialog now always starts with the first icon selected
- Fix wrong bounds for several typs of custom graphics in dialogs
- Fix wrong /source/ bounds for custom 28x36 graphics /everywhere in the game/.
- Fix select PC graphic dialog having a second page with an invalid graphic that can be selected.
- Warnings completely suppressed for the included TinyXML and gzstream libraries
- Parentheses warnings are now errors, since there were several that looked like bugs
- Ditto for dangling else warnings
Some of these warnings were actually bugs:
- Town wandering monsters would have never spawned, because the code to do so was accidentally nested within a check for overall_mode == MODE_OUTDOORS
---> boe.monster.cpp, lines 105-137
- Monster's behaviour with respect to elemental fields did not correctly depend on their immunities (this is the same precedence issue Sylae messed up fixing in the Windows code)
---> boe.monsters.cpp, lines 345-359
- Display of damage blocked by armour appeared to be incorrect (needs verification)
---> boe.newgraph.cpp, line 1079
- Three-choice dialogs probably weren't dealing with unusual button types correctly, though that's a minor point since they aren't expected to use such buttons
--> At least, it parses the sample file correctly
- I reverted the choice of using a skip parser, because it was failing and I couldn't figure out why. This means there's a lot of *ws where whitespace can go.
- Grammar rearranged a little
